Fix successful decompress reported as Error::Read #1612
Conversation
Streams less than 4096 bytes are sometimes reported as failed reads because stream_.avail_in is not reduced to 0. The next iteration of the loop finds `prev_avail_in == strm_.avail_in` and return false. `ret = inflate(...)` returns Z_STREAM_END on the first iteration of the loop indicating that inflate is finished. This fix prevents the second iteration of the loop from failing.
|
@bdenhollander could you please include at least one unit test in test/test.cc as I typically request from contributors when they send me a pull request? Thank you! |
- Add unit tests for raw deflate that illustrates the decompression failure when there are extra trailing bytes
|
@yhirose Writing the unit test helped narrow down the original source of the problem in my code that uses cpp-httplib, an incorrect Content-Length was including garbage after the end of the data I had compressed. Fetching the file with the extra bytes via browser fetch was successfully skipping the extra bytes but cpp-httplib couldn't handle them. |
Does your fix handle the above issue as well, right? |
Yes, that's correct. The second unit test recreates the scenario where there are extra bytes after the end of the Deflate compressed data and passes on this branch. The same test on the |
|
@bdenhollander thanks for your fine contribution! |
2 participants
Streams less than 4096 bytes are sometimes reported as failed reads because stream_.avail_in is not reduced to 0. The next iteration of the loop finds
prev_avail_in == strm_.avail_inand return false.ret = inflate(...)returnsZ_STREAM_ENDon the first iteration of the loop indicating that inflate is finished. This fix prevents the second iteration of the loop from failing.Relevant zlib.h source code:
https://github.com/madler/zlib/blob/04f42ceca40f73e2978b50e93806c2a18c1281fc/zlib.h#L462-L464